Clustering of multivariate geostatistical data

Abstract

AbstractMultivariate data indexed by spatial coordinates have become increasingly popular in many geoscientific fields. This type of data imposes new analysis challenges, including the clustering of irregularly sampled data locations into spatially contiguous clusters given a set of regionalized variables. Clusters of data locations created through general‐purpose clustering techniques turn out to show poor spatial contiguity, a characteristic obviously inconvenient for many geoscience applications. This article reviews clustering methods designed explicitly for multivariate geostatistical data in which spatial dependency plays an important role. These clustering techniques are modifications of general‐purpose clustering methods. They provide spatially contiguous clusters by accounting for the spatial dependency of data locations.This article is categorized under: Statistical Learning and Exploratory Methods of the Data Sciences > Clustering and Classification Data: Types and Structure > Image and Spatial Data